Estimativa de parâmetros geoelétricos e gravimétricos através de inversão híbrida com os métodos Metropolis e Steepest Descent.

Abstract

The main objective of this dissertation was to test inversion methodologies applied to synthetic gravimetric and geoelectrical data, integrating the inversion and interpretation of these data for the study of a hydrogeological environment. Based on the theory of potential fields, both methods face inherent ambiguities in the interpretation of their anomalies, which can be caused by several possible sources. To mitigate these limitations, joint inversion was used, which simultaneously processes the data, generating models that represent the geometry of the density and resistivity interfaces and the distribution of these properties. The study was conducted in two main stages: the individual modeling of the geophysical data and the application of two joint inversions, a global one, using the Metropolis method, and a local one, using the Steepest Descent, both implemented through codes developed in Python. Four initial models were evaluated under three noise levels (no noise, 5% and 10%), generating 24 models using the Metropolis methodology, which served as input for the Steepest Descent, totaling 48 inverted models. The results allowed us to evaluate the quality of the inversion methodologies, identify the limitations of each geophysical method and visualize the geometry and distribution of properties in the geological environment, contributing to the understanding and development of more robust solutions for inverse problems.